当数字说话：在文本到视频扩散模型中对齐文本数字和视觉实例
When Numbers Speak: Aligning Textual Numerals and Visual Instances in Text-to-Video Diffusion Models
AI总结：
本文提出NUMINA框架，通过识别和引导提升文本到视频扩散模型中数字对齐的准确性，实验显示在不同模型规模上均提升了计数精度，并保持了CLIP对齐和时间一致性。
摘要：
文本到视频扩散模型已能实现开放式的视频合成，但通常在生成提示中指定的物体数量时存在困难。我们引入NUMINA，一种无需训练的识别-引导框架，用于改进数字对齐。NUMINA通过选择判别性的自注意力和交叉注意力头来识别提示布局不一致，从而推导出可计数的潜在布局。然后，它保守地细化此布局并调节交叉注意力以引导再生。在引入的CountBench上，NUMINA在Wan2.1-1.3B模型上将计数准确性提高了7.4%，在5B和14B模型上分别提高了4.9%和5.5%。此外，CLIP对齐得到改善，同时保持了时间一致性。这些结果表明，结构引导补充了种子搜索和提示增强，为实现准确的文本到视频扩散提供了实用路径。代码可在https://github.com/H-EmbodVis/NUMINA上获得。
英文摘要：
Text-to-video diffusion models have enabled open-ended video synthesis, but often struggle with generating the correct number of objects specified in a prompt. We introduce NUMINA , a training-free identify-then-guide framework for improved numerical alignment. NUMINA identifies prompt-layout inconsistencies by selecting discriminative self- and cross-attention heads to derive a countable latent layout. It then refines this layout conservatively and modulates cross-attention to guide regeneration. On the introduced CountBench, NUMINA improves counting accuracy by up to 7.4% on Wan2.1-1.3B, and by 4.9% and 5.5% on 5B and 14B models, respectively. Furthermore, CLIP alignment is improved while maintaining temporal consistency. These results demonstrate that structural guidance complements seed search and prompt enhancement, offering a practical path toward count-accurate text-to-video diffusion. The code is available at https://github.com/H-EmbodVis/NUMINA.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
